Revisiting Water Injection Post-Polymer Flooding: A Global Perspective on Project Performance

Sagyndikov M., Dupuis G., Seright R., Thomas A., Wilton R. – 2025

Abstract

A major decision point in every polymer flood is what happens when polymer injection stops and water injection resumes. This transition is widely modelled as a continuation of the flood — often with the expectation that residual resistance factors and polymer-induced conformance improvements will sustain oil production. Field evidence suggests the opposite: recovery efficiency declines rapidly, water cut spikes, and production rates quickly return to pre-polymer baseline levels.

This paper synthesizes theoretical frameworks and field case studies from Kazakhstan, China, India, Oman, Brazil, and Canada to document the dynamics of post-polymer water injection and explain why standard simulation approaches consistently overestimate recovery during this phase. Root causes examined include the role of viscous fingering, the overestimation of residual resistance factors in high-permeability rock, the failure of graded viscosity bank models to account for real reservoir heterogeneity, and the tendency for water to re-establish high-permeability preferential flow paths that dominated before polymer injection. Practical recommendations are developed for extending the injection phase and designing polymer banks with post-flood performance in mind.

Key Takeaways

  • When water injection resumes after a polymer flood, production rapidly reverts to waterflood baseline — field cases from Kalamkas (Kazakhstan), Daqing (China), Mangala (India), Marmul (Oman), Carmopolis (Brazil), and Canadian heavy oil fields all show the same pattern: water cut spikes within days to weeks and oil rates return to pre-polymer levels.
  • Residual resistance factors in reservoirs above 200 mD should be assumed to be 1 for design purposes — reported high values in the literature are usually artefacts of insufficient post-polymer brine flushing during core experiments, and relying on them leads to overoptimistic post-flood forecasts.
  • Standard reservoir simulators consistently overestimate post-polymer waterflood performance because they fail to capture viscous fingering of chase water through the polymer bank in high-permeability pathways — the mechanism that most rapidly destroys polymer bank integrity.
  • Graded viscosity bank models, while theoretically appealing, are based on simplified 1D homogeneous systems. In heterogeneous reservoirs, water preferentially channels through the highest-permeability zones regardless of tapering strategy, and the benefit of grading is largely eliminated.
  • The clearest operational conclusion from the global dataset is that polymer injection should continue for as long as it is economically viable — switching back to water is effectively irreversible in most cases, and even temporary interruptions (Water-Alternating-Polymer effects) cause measurable recovery losses.
  • Primary and secondary polymer floods (without prior waterflood) consistently outperform tertiary polymer floods: in the Brintnell Wabiskaw field (Canada), early polymer floods produced most of their oil at a water-oil ratio of 0.1–1, versus 5–10 for tertiary floods, with 10% OOIP more incremental recovery before WOR reaches 1.
Read Article

Related Articles